Outcomes in Women and Minorities Compared With White Men 1 Year After Everolimus-Eluting Stent Implantation: Insights and Results From the PLATINUM Diversity and PROMUS Element Plus Post-Approval Study Pooled Analysis

Key Points

Question  Are 1-year outcomes among women and minorities different from white men after contemporary percutaneous coronary intervention, and what are the predictors?

Findings  In this pooled analysis study, women and minorities had more baseline comorbidities. After risk adjustment, 1-year major adverse cardiovascular events (death/myocardial infarction/target vessel revascularization), were not significantly different between groups; however, women and minorities had a higher adjusted risk of death/myocardial infarction than white men primarily because of a greater hazard of nonstent-related myocardial infarctions.

Meaning  Contemporary everolimus-eluting stents confer similar risks of major adverse cardiovascular events across race/ethnicity and sex; however, there remains opportunity to reduce an incremental risk of recurrent nonstent-related ischemic events in women and minority patients after percutaneous coronary intervention.

Importance  There exist limited outcomes data for women and minorities after contemporary percutaneous coronary intervention (PCI).

Objective  To examine 1-year outcomes in women and minorities vs white men after PCI with everolimus-eluting stents.

Design, Settings, and Participants  The PLATINUM Diversity study was a single-arm study enrolling women and minorities. Patient-level pooling with the PROMUS Element Plus Post-Approval Study was prespecified. Data on social determinants of health and language were collected in the PLATINUM Diversity cohort, which included 1501 patients at 52 US sites. The PROMUS Element Plus Post-Approval study enrolled 2681 patients at 52 US sites with some site overlap and included an “all-comers” population. All patients were enrolled beginning in October 2014 and were followed for 12 months. Analyses began in August 2016.

Interventions  Patients received 1 or more everolimus-eluting stent implantation.

Main Outcomes and Measures  The primary end point was 1-year major adverse cardiac events (MACE), which included death/myocardial infarction (MI)/target vessel revascularization. Secondary ischemic end points were also evaluated.

Results  The pooled study consisted of 4182 patients: 1635 white men (39.1%), 1863 women (white and minority) (44.5%), and 1059 minority patients (women and men) (25.3%). Women and minorities had a higher prevalence of diabetes, prior stroke, hypertension, renal disease, and congestive heart failure than white men but lower rates of multivessel disease, prior coronary artery bypass graft surgery, prior MI, and smoking. Unadjusted 1-year MACE rates (white men, 7.6%; women, 8.6%; minorities, 9.6%) were similar between groups with no significant differences after risk adjustment. The adjusted risk of death/MI was higher among women (odds ratio, 1.6; 95% CI, 1.1-2.4) and minorities (odds ratio, 1.9; 95% CI, 1.2-2.8) compared with white men and the adjusted risk of MI was higher in minorities (odds ratio, 2.6; 95% CI, 1.4-4.8). These differences were driven primarily by nonstent-related MIs. Within the PLATINUM Diversity cohort, the independent predictors of MACE were cardiogenic shock, renal disease, history of peripheral vascular disease, multivessel disease, widowhood, and lack of private insurance.

Conclusions and Relevance  After contemporary everolimus-eluting stent implantation, women and minorities experience a similar risk of 1-year MACE but a higher adjusted risk of recurrent ischemic events primarily because of nonstent-related MIs. Both clinical and angiographic factors and social determinants of health, including widowhood and insurance status, contribute to 1-year MACE among women and minorities.

With minorities expected to represent more than half of the population by 2060, the United States is becoming increasingly diverse.¹ Despite these trends, cardiovascular clinical trials have continued to include predominantly white men.² This enrollment bias persists despite major efforts from the US Food and Drug Administration and the National Institutes of Health to improve clinical trial diversity.^2-4 Given that cardiovascular disease remains the leading cause of death in men and women regardless of race/ethnicity,⁵ broad clinical trial inclusion is pivotal to providing an adequate evidence base from which new therapies may be assessed.

Women and minorities often present with more comorbidities than white men, thereby potentially contributing to differential outcomes.^6-8 Some, but not all, studies have shown a higher likelihood of postpercutaneous coronary intervention (PCI) adverse events in women, while others have reported higher risks in minorities.^8-14 However, these studies have yielded conflicting results and have been limited by outdated stent technologies, small sample sizes, short duration of follow-up, retrospective study designs, or have been devoid of angiographic data and/or insight into the effect of social determinants of health (SDOH).

The purpose of our study is to compare the baseline clinical, angiographic, and procedural characteristics and 1-year contemporary PCI outcomes of women and minorities with that of white men and explore the effect of SDOH.

The PLATINUM Diversity Study (PD) was a prospective, multicenter, open-label observational study that enrolled patients who received 1 or more everolimus-eluting stent and self-identified as having at least 1 of the following characteristics: female sex, black (of African heritage), Hispanic/Latino, or American Indian/Alaskan Native. There were no exclusion criteria. Patients were enrolled at 52 US sites beginning in October 2014 and followed for 12 months. Social determinants of health, including access to and use of health care, living circumstances, socioeconomic status, exercise habits, alcohol and drug use, and language barriers were collected prospectively in the PD study.

The PROMUS Element Plus Post-Approval Study (PE Plus) was a prospective, multicenter, open-label observational study completed in August 2014 that enrolled an “all-comers” cohort of patients across 52 US sites who were treated with an everolimus-eluting stent.¹⁵

The PD and PE Plus studies complied with the Declaration of Helsinki¹⁶ and were approved by each site’s locally appointed ethics committee. Written informed consent was required within 24 hours of stent implantation within both studies. Analyses began in August 2016. The PD and PE Plus studies are registered at http://www.clinicaltrials.gov under identifiers NCT02240810 and NCT01589978, respectively.

The everolimus-eluting stents used in the PD and PE Plus studies are thin-strut platinum chromium everolimus-eluting stents with a permanent polyvinylidene fluoride-co-hexafluoropropene polymer. Some of these stents had additional connectors within the proximal stent segments. These stents have demonstrated consistent safety and performance across the PD clinical trial program and in PE Plus .^15,17,18

The primary end point was the incidence of 1-year major adverse cardiovascular events (MACE), defined as a composite of all-cause death, myocardial infarction (MI), and target vessel revascularization (TVR). Additional secondary end points analyzed at 1 year included death (cardiac and noncardiac), death/MI, MI (stent-related and nonstent-related), Academic Research Consortium definite/probable stent thrombosis, and target vessel failure. All end points were adjudicated by an independent clinical events committee using established definitions from the PD and PE Plus studies.^15,17,18

Outcomes were compared between women vs white men and minorities vs white men, respectively. Two-sided t tests were used to compare continuous variables and χ² or Fisher exact tests were used for discrete variables. A χ² test compared the unadjusted outcomes between minority patients (black, Hispanic/Latino, and American Indian/Alaskan native) from the pooled cohort and white men from PE Plus. A similar method compared women from the pooled cohort and white men from PE Plus. Statistical significance was declared if the 2-sided lower 95% CI boundary on the difference between groups was less than 0 and the χ²P value was less than .05. Given the 284 minority and 806 female patients enrolled in PE Plus and a 7.3% 1-year MACE rate noted in the 1635 white men from PE Plus, we estimated that we would need 1500 women and minority patients in PD (assuming 5% attrition) to achieve more than 80% power to detect at least a 3% difference in MACEs. An assessment of the poolability of patients across PD and PE Plus sites was made by fitting a logistic regression model (eTable 1 in the Supplement).

Odds ratios (OR) and 95% CIs for the adjusted risk of 1-year MACE were generated using multivariate logistic regression. Baseline clinical, angiographic, and procedural characteristics were entered in the model (Figure). Covariates with a univariate P value of less than .10 were selected to remain in the model. To risk adjust, a group variable (white men vs women or vs minority patients) was forced into the model and the OR determined.

We first built a multivariate model predicting MACE from the entire pooled study sample. For women and minorities enrolled only in the PD cohort, we also created a separate model for MACE, including the SDOH data that were collected in this group. All statistical analyses were performed using SAS software, version 9.2 or later (SAS Institute Inc).

The study consisted of 4182 patients, including 1501 enrolled in the PD study and 2681 from the PE Plus study. The distribution of patients by race/ethnicity and sex and patient flow are shown in eFigure 1 and eFigure2 in the Supplement. The final pooled sample included 1635 white men (39.1%), 1863 women (white and minority) (44.5%), and 1059 minority patients (women and men) (25.3%).

Baseline characteristics are shown in Table 1. Compared with white men, women were older and had a higher prevalence of diabetes, hypertension, history of congestive heart failure, prior transient ischemic attack/cerebrovascular accident, renal disease, and lower rates of prior MI, multivessel disease, prior PCI, and prior coronary artery bypass graft surgery. Minority patients were slightly younger and had a higher prevalence of diabetes, hypertension, prior congestive heart failure, prior cerebrovascular accident, and renal disease and a lower prevalence of hyperlipidemia, family history of coronary artery disease, prior MI, prior coronary artery bypass graft surgery, atrial fibrillation, and multivessel disease. The prevalence of diabetes was highest among minority patients (535 of 1059 [50.5%]) followed by women (790 of 1863 [42.4%]) and white men (561 of 1635 [34.3%]). Minorities had nearly half the rate of preexisting atrial fibrillation (54 of 1059 [5.1%]) as white men (163 of 1635 [10.0%]) and women (171 of 1863 [9.2%]). Women and minorities were less likely to experience angina and women less likely to have silent ischemia than white men. Mean left ventricular ejection fraction was normal (>50%) in all groups, although slightly higher in women and slightly lower in minorities compared with white men.

Women and minorities had slightly smaller reference vessel diameter and a higher prevalence of moderate-to-severe lesion calcification and chronic total occlusions compared with white men. Women and white men shared similar lesion and stent lengths, with slightly longer lengths observed in minority patients. The prevalence of bifurcation lesions was highest in white men, whereas women had the highest prevalence of ostial lesions. Intravascular ultrasonography before stenting was performed at a similar rate across groups; however, intravascular ultrasonography after stenting was used more frequently among women and minorities. The use of fractional flow reserve testing was highest in minority patients. Women and minorities were treated with postdilation less often than white men. There were no differences between groups in the rate of urgent/emergent interventions and the number of lesions treated per patient.

The rate of aspirin and dual antiplatelet therapy (DAPT) use was slightly but significantly lower in women and minorities compared with white men (eFigure 3 in the Supplement). Dual antiplatelet therapy use at 6 and 12 months was 91% and 87%, respectively, for white men; 88% and 85%, respectively, for women; and 86% and 84%, respectively, for minority patients.

Social determinants of health for the PD cohort are shown in Table 2. Because the 2 groups (women and minorities) were not entirely mutually exclusive, P values for comparisons are not provided. Overall, women were older and more likely to have health care insurance and established primary care. For both groups, travel distance and time to the hospital performing the index PCI was significant, although this appeared to be more evident in women, such that nearly 404 of 1501 (26.9%) lived more than 48 km from the hospital. Women were more often retired, widowed, and better educated, while minorities were more often single, either currently employed or unemployed, and not retired. Of 1501 patients, 1056 (70.3%) were living with family. Although rates of nondisclosure of income were high, low income levels were common; 373 of all PD patients (24.9%) reported a household income of less than $25 000. Minorities were more likely than women to possess a smartphone, exercise regularly, and admit to alcohol and substance abuse. Minorities were also more likely to experience patient-physician language discordance.

The unadjusted primary 1-year MACE rate was similar across groups (white men, 7.6% vs women, 8.6%; [P = .33]; white men vs minorities, 9.6% [P = .08]) (Table 3 and eFigure 4 in the Supplement). Unadjusted rates of death/MI, death, and MI were higher in minority patients compared with white men (death/MI, 6.4% vs 3.1% [P = <.001]; death, 3.7% vs 2.2% [P = .03]; MI, 3.1% vs 1.1% [P = <.001]). Analogously, women had a higher unadjusted rate of death/MI and death than white men (death/MI, 5.0% vs 3.1% [P = .004]; death, 3.4% vs 2.2% [P = .04]) and a trend toward a higher rate of MI (1.9% vs 1.1% [P = .06]). Unadjusted rates of TVR, target vessel failure, and Academic Research Consortium stent thrombosis were similar across all 3 groups.

After multivariate adjustment for imbalances in baseline characteristics, there were no between-group differences in 1-year MACE, death, and TVR (Figure). The risk adjusted rate of death/MI was higher in both women (OR, 1.6; 95% CI, 1.11-2.39; P = .01) and minorities (OR, 1.9; 95% CI, 1.22-2.80; P = .004) compared with white men. Minorities also had a higher adjusted risk of MI (OR, 2.6; 95% CI, 1.40-4.82; P = .002), whereas in women there was only a trend toward a higher risk of MI (OR, 1.7; 95% CI, 0.91-3.09; P = .09). These differences were driven mostly by higher rates of nonstent-related MIs than stent-related events (Table 3). Compared with white men, there was a 2.6- and 5.3-fold higher risk of nonstent-related MI in women and minorities, respectively, with a 1.8-fold higher risk of stent-related MI seen only in minority patients.

The independent predictors of MACE for the entire pooled cohort were renal disease, prior cerebrovascular accident, multivessel disease, prior PCI, history of congestive heart failure, prior MI, and age (Table 4). For the PD cohort (women and minorities) in which SDOH were available, the independent predictors of MACE were cardiogenic shock, renal disease, peripheral vascular disease, multivessel disease, widowed status (vs married or divorced), and lack of private insurance (Table 4).

This study was specifically designed to prospectively collect comprehensive PCI outcomes data on women and minorities and compare with white men. We report that (1) with concerted effort and appropriate site selection, women and minorities may be rapidly enrolled in a large, multicenter, national PCI outcomes study with robust follow-up; (2) white men, women, and minorities share comparable adjusted risks of 1-year MACE, death, and TVR; (3) women and minorities appear to encounter higher risks of recurrent ischemic events (death/MI for women and death/MI and MI for minorities) that are driven primarily by nonstent-related MIs and; (4) SDOH, such as marital/widowed status and insurance status, appear to significantly affect outcomes in women and minorities.

Historically, white men have served as the prototypical clinical trial patient with outcomes then extrapolated to others.¹⁹ Because women and minorities may respond differently to medical or device interventions, building an adequate database for these underrepresented groups is germane. Impediments to achieving diverse study enrollment include participating site characteristics/location, broad accessibility of women and minorities to trial sites and investigators, willingness of patients to participate, and experience of the research team in obtaining informed consent, especially when English is not the first language.²⁰ To our knowledge, during the last decade, efforts to curb these trends have been unsuccessful.^6,21-23 For years, the US Food and Drug Administration has promoted the collection and presentation of data on race/ethnicity and sex in clinical trials.^3,23 Section 907 in the Food and Drug Administration Safety and Innovation Act of 2012²⁴ resulted in further efforts to refine the conduct and analysis of clinical trials to provide the best estimates of treatment effects in diverse populations. Similarly, the National Institutes of Health requires all funded clinical trials to include adequate enrollment of women and minorities.⁴ Still, only about a third of cardiovascular clinical trials collect data on race/ethnicity and sex, and, when reported, only 25% to 30% of US study enrollees are women and fewer are minorities.^2,25

Using an enriched cohort design for the PD study and pooling with the PE Plus cohort, our study included approximately 50% women and double the usual proportion of minorities. Until existing enrollment biases are eliminated, novel study designs, such as the present study, will be necessary to construct clinical databases more representative of US demographics. Counter to prevailing beliefs and a report suggesting that nonwhite race and patient-physician language discordance serve as impediments to clinical trial enrollment,²⁶ we demonstrate that, with appropriate site selection and concerted effort, women and minorities can be enrolled in a large, national, multicenter PCI outcomes study at a rate comparable with conventional studies. By completing enrollment 6 months ahead of schedule, the PD study enrolled patients faster than PE Plus. We attribute this to the highly motivated nature of our PD sites and their efforts to recruit these patients.

Consistent with prior studies, we observed that women and minorities present with more baseline comorbidities.^8-11,27-29 This likely accounts for the higher unadjusted rates of death and MI in these cohorts. However, after risk adjustment, MACE, death, and TVR were comparable between all 3 groups, suggesting that in the era of second generation everolimus-eluting stents, race/ethnicity and sex appear to have little to no effect on these outcomes. These findings contrast with those observed with first generation drug-eluting stents, for which worse outcomes, including death/MI and stent thrombosis, have been reported in black patients.^12,13,30 We suspect that advances in stent design (thinner struts and more biocompatible polymers) play an important role in producing uniformly more favorable stent outcomes across race/ethnicity and sex in our study. Despite this, we must acknowledge the higher adjusted risk of death/MI in women and minorities (and MI in minorities) compared with white men. This appeared to be mostly driven by a higher incidence of nonstent-related MIs which, in combination with similar TVR and stent thrombosis rates across groups, argues that stent failure itself contributed minimally to this incremental risk. We suspect that this phenomenon may be instead due to (1) race/ethnicity or sex–related factors that might affect thrombosis or atherosclerosis³¹; (2) the inability of our model to completely account for measured and/or unmeasured differences in baseline characteristics, including SDOH and; (3) differences in adherence to DAPT. Premature DAPT cessation has been linked to ischemic events.³² Although DAPT use at 6 and 12 months was lower in women and minorities than white men, the small differences observed (2.6% and 3.8% less DAPT use, respectively) probably do not solely account for the 60% and 90% increases, respectively, in the adjusted risk of death/MI. We hypothesize that observed lower DAPT use in women and minorities was due to a higher perceived risk of bleeding and/or concerns about patient nonadherence or cost. Despite the general reassurance that MACE was comparable across all groups, this observed incremental risk of nonstent-related ischemic events in women and minorities was intriguing and merits further investigation.

In addition to traditional clinical and angiographic risk factors, our study evaluated the incremental effect of SDOH on outcomes among women and minorities. The observation that cardiogenic shock, renal disease, peripheral vascular disease, and multivessel disease predicted MACE is in keeping with prior studies.^33-40 However, SDOH also added incrementally to the predictive model. We found that being widowed (vs married or divorced) increased the risk of MACE, while having private insurance was associated with a protective effect. Of the PD patients, 50% were married, 17% divorced, 16% single, 13% were widowed, and 2.9% did not disclose this information. Compared with being married or divorced, widowhood conferred a 1.9- and 2.6-fold incremental risk of MACE, respectively. There was also a 1.4-fold increased risk of MACE in widowed patients compared with those who were single; however, this did not reach statistical significance and therefore did not remain in the model. Widowhood has also been shown to increase mortality from heart disease and multiple other causes.^41,42 Major grief reactions, emotional stress, depression, and the loss of support associated with losing a spouse or partner may all contribute.⁴¹ Prior studies have also shown that insurance status may affect cardiovascular outcomes. Similar to our study, Gaglia et al⁴³ reported lower MACE rates in patients under age 65 years who had private insurance compared with Medicaid, Medicare, or no insurance. These findings and ours suggest that clinical and angiographic factors and SDOH may all affect post-PCI outcomes. Accordingly, we encourage future PCI outcomes studies to collect and report data on SDOH in addition to conventional PCI risk factors.

There were other insights gained from studying SDOH. Compared with women, minorities had lower degrees of education, lower annual household income, and lower rates of having an established primary care physician. Minorities were also more likely to have a foreign language as their primary language and to experience language discordance between themselves and their physicians. Although minorities reported regular exercise more often than women, paradoxically they also had higher rates of obesity. Prior studies have shown that education, health literacy, language barriers, and insurance status all play an important role in patients’ compliance and outcome.^44,45 Distance and accessibility to health care may also impact a patient’s choice of health care provider and outcome.^46,47 In our study, even though minority patients often lived less than 30 minutes away from the treating hospital, these patients had slightly lower follow-up rates. The fairly common use of smartphones suggests that digital technology may be a useful tool to improve compliance and follow up.^48-50 Insights from SDOH data may help clinicians formulate more race/ethnicity- and sex–specific programs to promote health, improve compliance, and reduce risk.

As patients were pooled from 2 separate PCI studies (PD and PE Plus), we cannot exclude the effects of unrecognized biases. We felt it valid to combine them because all patients were treated in a similar timeframe with comparable platinum-chromium everolimus-eluting stents, both studies enrolled consecutive US patients across 52 sites, and outcomes were adjudicated at the same time by an independent clinical events committee and using identical end point definitions. Furthermore, our logistic regression model provided additional statistical validity for pooling (eTable 1 in the Supplement). The overall follow-up rate at 1 year was high (92%-95%) but not perfect. However, follow-up was still robust across groups and in keeping with other contemporary PCI studies. Information on SDOH and medication compliance was self-reported and therefore prone to recall bias. Specific details on diet and exercise were not collected. Our study was statistically powered for MACE but not for other secondary end points, raising the potential for beta error, especially for infrequently encountered events such as stent thrombosis.

Women and minorities undergoing contemporary PCI present with more comorbidities than white men but share a similar adjusted risk of 1-year MACE, death, and TVR. Women and minorities experience a greater adjusted risk of 1-year death/MI (and MI for minorities), driven predominantly by a higher risk of nonstent-related MI. Our data suggest that differential risks of thrombosis or atherosclerosis contribute more to this phenomenon than stent failure. The independent predictors of MACE in women and minorities were cardiogenic shock, renal disease, peripheral vascular disease, multivessel disease, widowhood, and lack of private insurance. In providing a comprehensive evaluation of the impact of race/ethnicity, sex, and SDOH on contemporary PCI outcomes, this study contributes significantly to our current knowledge base.

Corresponding Author: Wayne Batchelor, MD, MHS, Florida State University College of Medicine, Southern Medical Group, P.A., 1300 Medical Dr, Tallahassee, FL 32308 (wabat@southern-med.com).

Accepted for Publication: August 30, 2017.

Published Online: October 18, 2017. doi:10.1001/jamacardio.2017.3802

Author Contributions: Drs Batchelor and Mehran had access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Batchelor and Mehran contributed equally to the work.

Concept and design: Batchelor, Singh, Underwood, Thompson, Mehran.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Batchelor, Davis, Singh, Lopez, Chandrasekhar, Underwood, Mehran.

Critical revision of the manuscript for important intellectual content: Batchelor, Kandzari, Tami, Wang, Othman, Gigliotti, Haghighat, Singh, Giugliano, Horwitz, Underwood, Thompson, Mehran.

Statistical analysis: Batchelor.

Obtained funding: Batchelor.

Administrative, technical, or material support: Othman, Lopez, Underwood, Thompson.

Supervision: Batchelor, Kandzari, Singh, Underwood, Thompson, Mehran.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Batchelor reports institutional grant/research support from Boston Scientific and consulting for Abbott, Medtronic, and Boston Scientific. Dr Kandzari reports minor consulting honoraria from Medtronic, Boston Scientific, and Micell Technologies and research/grant support from Medtronic, Boston Scientific, Biotronik, and St. Jude Medical/Abbott. Dr Wang reports consultant/executive committee for Boston Scientific. Dr Horwitz reports grant support from Boston Scientific. Drs Underwood and Thompson are full-time employees and stock holders of Boston Scientific Corporation.Dr Mehran reports institutional grant/research support from Daiichi-Sankyo/Eli Lilly, Bristol-Myers Squibb, AstraZeneca, The Medicines Company, OrbusNeich, Bayer, CSL Behring, Abbott Laboratories, Watermark Research Partners, Novartis Pharmaceuticals Medtronic, AUM Cardiovascular, Beth Israel Deaconess Medical Center; consultant/executive committee for Janssen Pharmaceuticals, Osprey Medical, Watermark Research Partners, Medscape, The Medicines Company, Boston Scientific, Merck, Cardiovascular Systems, Sanofi USA, Shanghai BraccoSine Pharmaceutical, AstraZeneca (all minor); and equity for Claret Medical and Elixir Medical Corporation.

Additional Contributions: We thank all of the PLATINUM Diversity and PROMUS Element Plus Post-Approval Study investigators and sites for their invaluable contributions to this work. We also thank Kristine Roy, PhD (Boston Scientific Corporation), for assistance in manuscript preparation and Songtao Jiang, MSc, and Chuyu Deng, BSc (Boston Scientific Corporation), for assistance with statistical analysis. Dr Roy and Mr Jiang are full-time employees and stock holders of Boston Scientific Corporation. Ms Deng was an intern and received a salary from Boston Scientific Corporation at the time of manuscript submission.